Monitoring here consists of three stock standard linux tools dstat, iostat, and pidstat I won’t even pretend that I know how to do this on Windows, but I imagine they have some equivalent, feel free to comment and educate me, but I don’t have the cycles to do the play around with Windows as well.

The dstat tool is a consolidated replacement for a few previous tools that did each task, so to simplify it is a good starting point to help us cover our system level monitoring, and pidstat. Let’s walk through some quick examples of each.

Here dstat is doing most of the heavy lifting and captures the time, cpu, memory, and network i/o with the defaul delay of every 1 seconds.

dstat --time --cpu --mem --net 1 > logs/dstat.log &

For pidstat we need to identify the process we want to look at first, then we can start watching it.

pid=`ps aux|grep -v grep|grep ibm-ucd/server|awk {'print $2'}`
nohup pidstat -I -p ${pid} -u -d -r -h -l 1 > logs/pidstat.log &

Reference:

• dstat
• iostat
• pidstat
• sysstat tools